> Continuing with your naive reasoning, let's make synthetic corn instead of farming too because why not needlessly maximize complexity of the systems that we have to maintain?

Sarcasm aside, it is worth exploring. The typical photosynthetic efficiency of crops is only in the 1-2% range.

Modern solar panels reach 20%. The electricity can be used to produce hydrogen at 80% efficiency. The hydrogen could be used as energy input for an engineered yeast to produce proteins that we need.

I don't know about the efficiency of that last step, but it is at least plausible that the overall process could be more efficient than photosynthesis. Solar Foods[1] is betting that it will be.

[1] https://en.wikipedia.org/wiki/Solar_Foods

That is actually very cool and a much more promising path than animal cell culture, but tbh even with 10+ years of sustained investment I have serious doubts about whether it would compete with lower yielding sustainable ag efficiency at scale. Glad someone is doing the research though.

Seems like there is a two step conversion, step 1 is 40-50% efficient in mass conversion H + co2 to acetate by Clostridium ljungdahlii, then a 25% efficiency conversion by yeast to biomass. So that's already down to 2% overall for .07g/L/h at lab scale. Then additional losses in down stream processing to remove the water. The output is more of a yeast protien meat replacement.

https://www.chemistryworld.com/news/microbes-and-renewable-e...

It's extremely, extremely unlikely that we're going to beat plants on efficiency of biomass production, on either a per unit solar energy or per unit resources basis (there are two general photosynthesis pathways that optimize for each of these two endpoints). That doesn't mean we can't beat them at other things, of course (as solar panels demonstrate) but this is what they are optimized for, and the competitive advantage of finding a better way to synthesize biomass is huge to the point that it only needs to evolve once to take over a large fraction of the Earth.

> It's extremely, extremely unlikely that we're going to beat plants on efficiency of biomass production

I'm not so sure this is true. If it turned out that a more efficient pathway is possible with solar energy collection via sheets of extremely pure crystalline silicon (i.e. solar panels), then I would not think that it's strange that evolution didn't come up with it first. Some solutions simply aren't accessible to evolution.

Update: a relevant article appeared on phys.org this week about starch synthesis.

https://phys.org/news/2021-09-chinese-scientists-starch-synt...

> "According to the current technical parameters, the annual production of starch in a one-cubic-meter bioreactor theoretically equates with the starch annual yield from growing 1/3 hectare of maize without considering the energy input," said Cai Tao, lead author of the study.

It sounds like they're trying to save on land and freshwater, not energy or raw materials.